PolySpace for C - Ada€¦ · PolySpace for C A new approach ... first automatic detection of...
2
PolySpace for C A new approach PolySpace allows you to detect run-time errors and concurrent accesses on shared data automatically very early in the development process without testing . It statically analyses the dynamics of software applications by relying solely on the source code. • No test cases to write; • No modification of the code; • No application execution necessary. In contrast to dynamic testing, PolySpace automatically and directly highlights causes of run-time errors in the source code (therefore, no debugging). The net result is a drastic reduction in software development costs by dramatically reducing the need for testing and debugging. Do you want to reduce your coding and testing effort? List of operations causing Runtime Errors with macro expansion and filters for quickly focusing on the most critical errors. Each operation checked is displayed using a meaningful colour scheme and related diagnostic: Red Errors which occur at every execution, Green Error conditions that will never occur, Orange Warning – an error may occur sometimes, Gray Shows unreachable (dead) code. Coloured source code using the PolySpace scheme. Diagnostic for each Runtime Error detected for a quick understanding of the errors. The industrial challenge Embedded systems are at the heart of the world’s most critical applications and assuring their reliability is of paramount concern. The increase in size and complexity of current applications results in rising testing costs. It also highlights the limitations of available tools and methods. Indeed, when it comes to finding runtime errors, classical methods are inadequate. For example, tools that compute metrics or enforce the use of coding rules may help in introducing fewer bugs during the coding but will not actually find bugs. Dynamic testing, which is of undisputable value for functional validation, only catches anomalies when they surface during test case execution. Furthermore, with traditional approaches, every error detected during program execution requires hours, days, even weeks of debugging. Today’s applications require a new generation testing solution that is capable of improving software quality while reducing the usual costs associated with such efforts. Runtime Errors detected Run-time errors are faults whose consequences include processor halt, data corruption or security breaches. They are latent faults that generally surface during execution. They have a major business impact: testing goes over budget, delayed releases, service disruptions during operation (e.g. sending uncontrolled commands to external devices). Errors detected by PolySpace include: • Read access to non-initialized data (variables and function return values); • De-referencing through null and out-of-bounds pointers; • Out-of-bounds array access; • Invalid arithmetic operations such as division by zero, sqrt(negative number); • Overflow / underflow on arithmetic operations for integers and floating point numbers; • Dangerous type conversions e.g. long → short, float → int; • Access conflicts for data shared between threads. PolySpace also detects: • Non-terminating function calls and loops; • Unreachable code (dead code).
-
Upload
duongkhanh -
Category
Documents
-
view
218 -
download
0
Transcript of PolySpace for C - Ada€¦ · PolySpace for C A new approach ... first automatic detection of...
PolySpace for C
A new approach
PolySpace allows you to detect run-time errors andconcurrent accesses on shared data automatically veryearly in the development process without testing.
It statically analyses the dynamics of software applicationsby relying solely on the source code.
• No test cases to write;• No modification of the code;• No application execution necessary.
In contrast to dynamic testing, PolySpace automatically anddirectly highlights causes of run-time errors in the source code(therefore, no debugging).
The net result is a drastic reduction in softwaredevelopment costs by dramatically reducing the need fortesting and debugging.
Do you want to reduce your coding and testing effort?
List of operations causing Runtime Errors with macro expansion and filters for quickly focusing on the mostcritical errors. Each operation checked is displayed using a meaningful colour scheme and related diagnostic:Red Errors which occur at every execution,Green Error conditions that will never occur,Orange Warning – an error may occur sometimes,Gray Shows unreachable (dead) code. Coloured source code using
the PolySpace scheme.
Diagnostic for each Runtime Error detectedfor a quick understanding of the errors.
The industrial challenge
Embedded systems are at the heart of the world’s mostcritical applications and assuring their reliability is ofparamount concern. The increase in size and complexity ofcurrent applications results in rising testing costs. It alsohighlights the limitations of available tools and methods.
Indeed, when it comes to finding runtime errors, classicalmethods are inadequate. For example, tools that computemetrics or enforce the use of coding rules may help inintroducing fewer bugs during the coding but will not actuallyfind bugs. Dynamic testing, which is of undisputable valuefor functional validation, only catches anomalies when theysurface during test case execution. Furthermore, withtraditional approaches, every error detected during programexecution requires hours, days, even weeks of debugging.
Today’s applications require a new generation testingsolution that is capable of improving software qualitywhile reducing the usual costs associated with such efforts.
Runtime Errors detected
Run-time errors are faults whose consequences includeprocessor halt, data corruption or security breaches.They are latent faults that generally surface duringexecution. They have a major business impact: testinggoes over budget, delayed releases, service disruptionsduring operation (e.g. sending uncontrolled commandsto external devices).
Errors detected by PolySpace include:• Read access to non-initialized data
(variables and function return values);• De-referencing through null
and out-of-bounds pointers;• Out-of-bounds array access; • Invalid arithmetic operations such as
division by zero, sqrt(negative number);• Overflow / underflow on arithmetic operations
for integers and floating point numbers;• Dangerous type conversions
e.g. long � short, float � int;• Access conflicts for data shared between threads.
PolySpace also detects:• Non-terminating function calls and loops;• Unreachable code (dead code).
Benefits
PolySpace brings both quality and productivity improvementsfor developers of embedded applications as well as for qualityassurance engineers.
PolySpace Developer EditionDevelopers benefit from an early detection of errors – beforeany testing – which is the key and crucial element to reducingtesting and debugging times. PolySpace offers the earliest andfirst automatic detection of run-time errors at compilationtime. Indeed, you just have to click on a source code file toget immediate results on your workstation. With no testcases to write, no instrumentation or execution of theapplication, PolySpace produces immediate savings. Theerrors that PolySpace finds early in development wouldinvolve appreciable debugging time if detected later duringthe test phase. Furthermore, performing functional tests ona module analysed by PolySpace becomes much easier thanksto more robust source code. So, why spend time debuggingruntime errors when PolySpace would have detected themearlier and automatically?
PolySpace Auditor EditionQuality Assurance Engineers get a reliable assessment of thequality of software applications, which is a key issue intoday’s supplier/end user relationship. PolySpace offers a wayto improve the quality of software applications automatically,by drastically reducing the number of bugs that can surfaceduring late system test after code freeze as well as at the end-users site. PolySpace provides fast, repeatable and low costacceptance criteria for software applications. It is suitable forboth internal quality control performed by an independentvalidation and verification team as well as to assess outsourceddevelopment and test activities.
Email: [email protected]
Static Verification
Static verification of dynamic properties is a well-establishedmethod for representing all possible executions of a givensoftware application, based solely on the source code. The basic idea is to calculate the domain of values eachvariable can take, for each point in the application by applyingmulti-thread, interprocedural and control structures analysis.This technique is used by PolySpace to identify run-timeerrors. This automatic approach makes PolySpace the uniquetesting tool that can statically predict which run-time errorswill affect your application prior to tests and deployment, e.g.
int tab [100] ;int *p = tab ;int i ;for (i = 0 ; i < 100 ; i++, p++) *p = 0 ;*p = 5 ;
The expression *p = 5 ; causes a memory corruption thatis automatically detected by PolySpace. All other testingmethods would request human efforts to catch it (codereview, test execution or code instrumentation).
Supported C Language Features & Software Engineering Standards
PolySpace supports C as defined in the ISO/IEC 9899 : 1990(E) standard. Furthermore PolySpace tools naturally integrate with softwareengineering standards such as MISRA, DO-178B, IEC 61508,CENELEC, FDA, TickIt, CMM, or IEEE 1012. They area means to automate or improve several activities prescribedby these standards such as code evaluations, static analysis,control/data flow analysis, providing definite transition criteriaand avoidance of runtime errors.
PolySpace finds more bugs than classical testing & much faster.
A complete andinteractive Call Tree
is provided fordocumentation and
navigation purposes.Fully exportable to
Excel © format.
Global Data Dictionary including read andwrite access to all global data (pointer and aliasincluded). It also identifies data shared betweenthreads and related protection mechanisms(critical sections, temporal exclusions).Fully exportable to Excel © format.
Access Graph for each Shared Data item:A meaningful and efficient feature showing how threads
access shared data through function calls
